Auto-Registration not Enabling the Host

[nvrohanraj]

Hi Everyone,

We have a client who monitors the Desktop/laptops/Tabs with Zabbix. This is a client requirement.
The issue is that we have once the Client Office Staff has left the office the laptop and tab is not reachable and the desktop is shutdown while leaving office.
In Zabbix the queue goes up once the devices are offline
I have disabled the host once the agent is offline but the auto registration is not reenabling it.
Nothing in the Server log. In the Proxy Log and the Agent Log it says the Host is not Monitored.
Auto-Registration Contains the operation to enable the host

Steps to reproduce:

1. Shutdown the Laptop that has the agent installed and working.
2. Waited for the Trigger of Host offline.
3. Disabled the Host in Zabbix.
4. Started the Laptop.
5. In the Agent Log it says the Host is not Monitored.

Changes in configuration...
No changes to the Config.


Result:
The Agent Log says the Host is not Monitored. The Proxy says not monitored

Expected:
Since the host is disabled the Auto-Registration should reenable the Host.

[MRedbourne]

You don't provide a Zabbix version - but regardless, this sounds like behaviour consistent with what I'd expect of Autoregistration.

Autoregistration is meant for registration, not re-enabling hosts that have been turned off. 2 Active agent autoregistration (zabbix.com). From the documentation (I referenced 6.0 LTS, but this is applicable for 6.4 as well), reference this line:

"[...] Autoregistration can happen when a previously unknown active agent asks for checks." In this case, your agent on the endpoint is previously known. Therefore, autoregistration is NOT being re-applied because it's not registering itself - it's already done that at setup.

What is your end goal? Are you looking to decrease the number of alarms you receive (via whatever media is setup), or are you attempting to disable the alert outright in your Zabbix web front end? If you're simply looking to stop receiving alerts from Zabbix for hosts that go "offline" during expected hours, you can look at creating a new 'reporting' user in Zabbix and define its media hours there. (Eg: email [email protected], M-F, 8AM-5PM), then create a new trigger action to report issues to that 'user', even if the same email is used between two different users. You'd need to split by Host Group or Host, most likely. (I'd recommend Host Group, you can dynamically add hosts into it using Auto Registration)

What I think you're actually looking for is a "maintenance period". (See: 11. Maintenance (zabbix.com)) You would define a maintenance period for these devices for non-working hours. Let's say you defne this as being between the hours of 5PM (1700hrs) and 8AM (0800hrs) M-F, and all of Saturday and Sunday. Give it a Host Group to use - which should be easily maintainnable for you given you're using Autoregistration. Set the "Maintenance" type to "No data collection" which will prevent triggers relying on nodata() from firing.



I hope this helps.

**I am not a Zabbix staff member. I'm simply an administrator for Zabbix for my company and a member of the broader Zabbix community.**

[nvrohanraj]

The reason I did not mention is because we have mostly 3 versions rather than just a specific version, Anyways The challenge is mainly related to the alerts and Queue. We have close to 150 Desktops and 150 Tabs like Microsoft Surface. The Queue becomes more then 1000 Because of which Zabbix does not retry some important checks.
If the Host is disabled once the host is offline probably a plan we thought and use the auto registration.
The issue we had is that once we disable the host, the alerts also hidden.

[MRedbourne]

OK,

I understand the role of the laptops and tablets (being user endpoints). What's the desktop in this case, the proxy monitoring the tablets/laptops? What I would do is create a host group (or two) for the client and stick the devices in those host group(s). With that, I'd put them into a maintenance period when outside of [their] outside of business hours. We have a test host (of sorts, it's not in production yet) where I tested this. It seems to work just fine. When you put these hosts into maintenance mode, any alerts should be suppressed until they're out of maintenance mode. Uncheck the "Show suppressed problems" section if it's generating alerts.



This was the maintenance period section:



This shouldn't affect your Zabbix item queue. At least, it's not affecting it on my side. (I do have a queue, but it's because one of my monitored hosts has some clockdrift that I need to go fix that's unrelated to this).